Led the development of a machine learning platform supporting the full model lifecycle—training, validation, deployment, and monitoring—for autonomous driving models. Built distributed training infrastructure on Kubernetes and Docker to handle large-scale deep learning jobs, improving research-to-production turnaround. Integrated ML compiler/runtime optimizations (ONNX Runtime, TensorRT) to reduce inference latency and memory footprint. Built backend APIs for model serving and telemetry data aggregation; implemented monitoring pipelines with ELK Stack and Prometheus; designed a React dashboard for operators to monitor fleet-wide performance. Guided a 12-member team of ML engineers and infra developers to deliver production ML systems on time with high reliability.

Senior Software Engineer – Identity Infrastructure at Google

April 1, 2021 - October 25, 2025

Designed and scaled Google's global authentication platform, processing 100M+ QPS across multiple data centers; built distributed caching and storage systems (60M+ QPS cache, Bigtable, GCS) to support ultra-low-latency identity lookups. Implemented microservices and gRPC-based inter-service communication; improved throughput and fault tolerance with automated failover and load balancing. Developed token-usage analytics pipelines with Apache Beam and Cloud Dataflow and maintained high availability with robust monitoring and incident response.

Team Lead – Spectrum Access System (SAS) at Google

April 1, 2017 - October 25, 2025

Led an 8-engineer team building the Spectrum Access System (SAS) for CBRS spectrum management; designed backend services on Kubernetes and GCP Compute Engine; built real-time ingestion with Apache Kafka; implemented distributed storage with PostgreSQL, Bigtable, and MongoDB; developed REST and GraphQL APIs and real-time dashboards; integrated OAuth/JWT-based identity management and caching for low-latency queries; led Agile processes and coordinated cross-functional collaboration.